Stanford University Gangnam Style flash mob
One question that has been argued endlessly about hit songs, movies, TV shows and books is whether there's something intrinsic that explains their success. Are hits simply of higher quality (or, to avoid qualitative judgments, do they better match consumer preferences) than less successful songs, movies, TV shows and books? Or are they popular because the mass culture industry spends billions of dollars annually to manufacture a demand for what is essentially interchangeable product?
If success is just a matter of knowing the existing market or jamming prefabricated product down our throats, the surprise is that success comes so infrequently. As Duncan Watts reports in his book Everything is Obvious (Once You Know The Answer) (Crown, 2011), the Beatles had trouble getting their first recording contract, George Lucas struggled to get financing for Star Wars, the Fox network turned down Friends, and JK Rowling's Harry Potter was rejected by eight publishers. Part of the problem is that our knowledge of the success of these cultural products is retrospective: each of them helped create a mass consumer demand for products like themselves. It's not clear that if you had exhaustively polled pop music fans in 1962 you would have come up with the Beatles as the ideal group, since very few people had ever heard them.
But there's another reason that it's so hard to predict success. Yes, there are differences among cultural products, and yes, the culture industry tries mightily to create demand. But especially in the instant-feedback age of the internet, a chief reason why certain products become popular is because, well, they're popular. It's a process that sociologists term "cumulative advantage": once a song or movie gains a slight popularity advantage over its rivals, that advantage tends to get amplified. Once enough of your friends, workmates or other information sources are mentioning this goofy Gangnam Style video, you have to see it yourself (particularly since the barriers of cost and inconvenience for doing so are minimal), and soon hundreds of millions of people have done likewise. We may also use information about other people's choices as a filter mechanism: faced with an overwhelming number of new novels, movies, and songs, we may be tempted to assume that there's a correlation between popularity and quality.
Watts and his research partners decided to try to measure the size of social influence on musical choice. We obviously can't re-run the early 1960s to see if the Beatles become popular again, so Watts and his collaborators came up with an ingenious solution. They created a website offering free downloads of 48 songs by new bands. What the visitors to the website didn't realize was that they were seeing one of 9 different versions of the site. On one version (the "independent world"), the songs were arranged randomly, and no information about other visitors' ratings or downloads was given. In eight other versions (the "social influence worlds") the number of downloads was shown. The researchers ran two experiments on the social influence worlds, one in which the songs were arranged randomly, and one in which the songs were listed in descending order by the number of downloads. Visitors could not download a song without listening to it.
So how much of an effect did social influence have on the songs visitors chose to listen to and download? A huge one. In the randomly-arranged independent world, the chances that any two songs would be listened to were equal, as we might expect. In the social influence worlds where the songs were arranged randomly, the visitors were three times more likely to listen to the most popular song than to songs of average popularity. And in the social influence worlds where the songs were ordered by the number of downloads, visitors were ten times more likely to listen to the most popular song than a song of average popularity.
And the popular songs were not always the same in each of the worlds. The researchers measured popularity by the number of downloads, and used the results from the independent world as establishing a baseline ranking of each song's appeal. While the songs with the most appeal generally did better in all the social influence worlds, the results for individual songs were highly unpredictable. Watts gives the example of the song "Lockdown" by 52 Metro, which ranked 26th out of 48 in appeal in the independent world. In one social influence world, it ended up as no. 1 in downloads, while in another it ended up as no. 40. Across all the worlds, a song in the top 5 in appeal had only a 50 percent chance of winding up among the top 5 downloads.
So this result suggests that if we could re-run the pop culture history of the past six months or so, there's a good chance that "Gangnam Style" would not have become a worldwide phenomenon. But then we wouldn't have had the pleasure (?) of watching the Stanford University Gangnam Style Flash Mob.
Watts, Duncan. (2012). Everything Is Obvious (Once You Know the Answer). New York: Crown Publishing. Chapter 3: The wisdom (and madness) of crowds, pp. 54-81.
Salganik, M. J. & Watts, D. J. (2009). Web-based experiments for the study of collective social dynamics in cultural markets. Topics in Cognitive Science 1, 439-468.
PSY: Gangnam Style (Official video)
The original dancing flash mob: Sound of Music | Central Station Antwerp